High intensity focused ultrasound (HIFU), applied in multiple, short pulses has great promise as a noninvasive therapy for several diseases, particularly cancer. At present, this and all other "High Temperature Therapy" heating modalities 1) deliver unknown and potentially highly variable, thermal dose distributions (resulting in patient pain, normal tissue damage and/or inadequate treatment of the clinical target) and/or 2) can require inordinate amounts of time (to deliver a complete, yet safe dose), particularly for "large" tumors. To overcome these limitations we propose to develop and comparatively evaluate a MR temperature based, optimization and feedback control approach for implementing such treatments, using HIFU as the heating modality. The optimized plans will serve as the feed-forward signal to a model predictive, thermal dose controller to vary the applied power magnitude (in both time and space) on-line during the treatment. Several recent research advances at the University of Utah make our proposed optimization/control approach particularly unique and promising, including the development of: 1) Model Predictive HTT Control techniques--the best control methodology to apply to these non-linear, distributed systems; 2) new, reduced order models (ROMs) for use in on-line, real time model based control, and; 3) Dynamic Inversion to help plan treatments. The comparative studies will concentrate on determining 1) the advantages that dose based optimization and control approaches have over the current, non-optimized, "model-less", ad hoc power, temperature and imaging control approaches and 2) the level of modeling needed for delivering high quality clinical treatments. Our newly developed MRA method for detecting, locating, and segmenting thermally significant blood vessels and our recent advances in thermal and ultrasound modeling will be used to construct accurate, detailed patient models for the comparative studies. Those studies will use simulations, dynamic phantoms and animal tests, and will initially concentrate on brain and sarcoma sites. The successful development and testing of the proposed approach will result in an overall system that will make HIFU optimally prescribable, controllable and evaluable (i.e. clinically practical), important features that do not exist at present. More generally, because of the basic nature of the optimization and control tools to be developed, these advances will be readily other HTT heating modalities.